基于支撑向量机的空瓶智能检测方法

为实现空瓶的智能检测,对空瓶检测的机器视觉方法进行了深入研究.论述了如何进行处理区域的标定,以及提取空瓶图像特征的方法.在提取了特征之后,提出用遗传支撑向量机算法来进行分类决策,支撑向量机具有较好的推广能力.同时采用遗传算法来优化选择支撑向量机的参数,以保证支撑向量机具有优良的分类性能.实验表明,采用这种方法检测空瓶,其检测准确率可达95%以上.

Empty Bottle Intelligent Inspecting Method Based on Support Vector Machines

The empty bottle inspected methods are discussed based on machine vision for inspecting empty bottles intelligently.A method is presented to identify the processing region,and extract features from images.After extracting feature,the GA-support vector machines(SVMs) are used as classifier.The SVMs have good generalizing ability.At the same time an optimization algorithm based on genetic algorithm is proposed to confirm the parameters of the support vector machines,and it ensure that the SVMs have good classification ability.The experiments show that by using this method to inspect empty bottles,the accuracy rate may reach above 95%.